National Development Strategy Croatia 2030 Policy Note: Macroeconomic Stability, Fiscal Policy and Taxation July 2019 Co-financed by the Technical Assistance of the Operational Program “Competitiveness and Cohesion” from the European Regional Dev elopment Fund Acknowledgements This policy note was prepared in the context of the Reimbursable Advisory Services Agreement “Sup- port for Establishing the System for Strategic Planning and Development Management and for Preparing the 2030 National Development Strategy”. The core World Bank team was led by Donato De Rosa (Lead Economist, Team Leader), Josip Funda (Senior Economist, co-Team Leader), and Catalin Pauna (former Team Leader) and included Stanka Crvik Oreskovic (Project Coordinator) and Bogdanka Krtinic (Program Assistant). The team worked under the guidance of Arup Banerji (Country Director), Elisabetta Capannelli (Country Manager) and Gallina Andronova Vincelette (Practice Manager). Preparation of the policy note Macroeconomic stability, fiscal policy and taxation was led by Josip Funda (Senior Economist). The main author of the policy note is Josip Funda with contributions from Darjan Milutinovic (Consultant), Alan Bobetko (Consultant) and Donato De Rosa (Lead Economist). Preparation of the policy note was guided by Gallina Andronova Vincelette as Practice Manager for Macroeconomics, Trade & Investment Global Practice for Europe and Central Asia, and Rogier J. E. van den Brink as a Program Leader. The policy note team thanks the following individuals and organizations in Croatia: • the Ministry of Regional Development and EU Funds for overall coordination and guidance, especially Ana Odak, the Assistant Minister, and her team; • the Ministry of Finance, especially Stipe Zupan, the Assistant Minister, and his team for meetings and consultations that have informed the policy note. Note This report is a product of the staff of the World Bank Group. The findings, interpretations, and conclu- sions expressed in the report do not necessarily reflect the views of The World Bank Group, its Board of Executive Directors, or the governments they represent. The World Bank Group does not guarantee the accuracy of the data included in this work, which is drawn from multiple external sources. Nothing herein shall constitute, or be considered to be, a limitation upon or waiver of the privileges and immun- ities of The World Bank Group, all of which are specifically reserved. Macroeconomic stability, fiscal policy and taxation 2 Contents Executive Summary ................................................................................................................................ 4 1 Introduction .......................................................................................................................................... 5 2 Macroeconomic Developments ............................................................................................................ 6 3 Potential Output.................................................................................................................................. 10 4 Exports of Goods and Services .......................................................................................................... 19 5 Public finances ................................................................................................................................... 27 5.1 General government deficit and debt ....................................................................................... 27 5.2 General government revenues .................................................................................................. 30 5.3 General government expenditures ............................................................................................ 33 6 Public Sector Performance and Efficiency ......................................................................................... 36 7 Conclusion .......................................................................................................................................... 41 List of references ................................................................................................................................... 42 Annex I. Public sector performance and efficiency: variables and methodology ................................. 43 1 Variables and series..................................................................................................................... 43 2 Methodology and data ................................................................................................................. 43 2.1 Public sector performance ........................................................................................................ 43 2.2 Public sector efficiency ............................................................................................................ 45 2.3 Data envelopment analysis ....................................................................................................... 46 Annex II. Debt Sustainability Analysis Assumptions ........................................................................... 47 Macroeconomic stability, fiscal policy and taxation 3 Executive Summary Croatia has made significant economic progress since independence, but the global financial crisis has exposed weaknesses in its economy and growth model. While growth has resumed in 2015, income convergence has been thrown back by the crisis and by the prolonged recession that followed and is currently proceeding only slowly. Low growth potential, deriving from low productivity growth, an ageing society, emigration, and a pace of capital accumulation lower than before the crisis, needs to be boosted for sustainable and rea- sonably rapid income convergence with the EU28. To this end, the export sector merits particular attention. Integration in global value chains and exposure to international competition would help to strengthen firm-level productivity, investment, and wages. Croatia has the smallest goods export sector among its peers in Central and Eastern Europe (CEE), and while exports of services are strong thanks to a vibrant tourism industry, the potential for productivity improvements there is limited. Fiscal policy will need to remain tight, as high public debt—to a significant extent denominated in foreign currency—needs to be reduced and fiscal space rebuilt. However, there is scope for using ex- penditure policies, for instance better targeting of social assistance, to increase incentives for labor force participation, which would help counter adverse demographic trends, reduce the need for fiscal transfers and increase the scope for raising public investment, and reduce inequality and poverty. Lastly, the effectiveness and efficiency of public sector service delivery across almost the entire public sector will need to be raised. This would also serve to make Croatia a more attractive destination for investments, which in turn would help boost output, reduce unemployment, and provide more reve- nue. Tackling these interrelated challenges requires a broad and well-coordinated reform agenda. Macroeconomic stability, fiscal policy and taxation 4 1 Introduction Croatia has achieved significant progress in reforms and living standards since independence. However, the 2008 global crisis has exposed its structural weaknesses and the unsustainability of its pre- crisis growth model. While growth resumed in 2015, substantial structural weaknesses remain, and Cro- atia’s real convergence process is yet to be restored at a pre-crisis pace. This paper, after an overview of macroeconomic developments over the past decade, discusses four interrelated issues crucial for accelerating income convergence with the EU28: I) Potential Growth. Raising Croatia’s growth potential is a sine qua non for income conver- gence, and would help in alleviating strains on the economy and public finances stemming from population ageing. Key to raising potential growth is boosting productivity. II) Exports. The export sector is an important engine for achieving higher productivity in the economy as a whole. Export-oriented firms generally pay higher wages, have higher sales, and higher rates of capital utilization, driven by exposure to foreign competition and integra- tion in Global (and European) Value Chains (GVCs). III) Public Finances. Fiscal risks have receded but remain sizeable. This necessitates continuation of tight fiscal policies. At the same time, the structure of revenues and expenditures could be strengthened to boost incentives for work, e.g. by better targeting of social assistance. This could also contribute to reducing inequality and poverty. IV) Public Sector Efficiency. Croatia’s public sector performance in almost all dimensions is lag- ging CEE peers in terms of effectiveness as well as efficiency. The state is large in terms of spending, but outcomes in health, education, administrative efficiency, distribution, stability and overall economic performance are weak. A broad reform agenda is needed to boost productivity and competitiveness, raise the quality of human and physical capital and modernize public services. While reforms need to be prioritized, it must be recognized that they are often interrelated and can reinforce each other. Macroeconomic stability, fiscal policy and taxation 5 2 Macroeconomic Developments Croatia’s pre-2008 growth model and income convergence were unsustainable, as they were largely debt financed, and accompanied by low productivity growth. The recovery since 2015 is broad-based, and the economy’s structure has improved. Nonetheless, income convergence proceeds only slowly, with pre-crisis levels not yet attained. Building on the market-oriented structural reforms of the 1990s and regained political stabil- ity, Croatia enjoyed a period of sustained economic growth and income convergence with the EU before the global economic crisis. During 2002-08, real GDP per capita rose annually by 4.5 percent on average, and its level more than doubled to above US$20,000 reaching 63 percent of the EU28 aver- age (at PPS).1 As a result, Croatia’s real convergence2 with EU28 income levels was similar to countries in Central Europe but slower than that of countries that started from a lower level of development (e.g. Baltic countries and Romania) (Figure 1).3 During this time, growth was mainly driven by domestic demand, fueled by large capital in- flows and robust credit growth. The main drivers of expansion were capital accumulation and private consumption (Figure 2). Gross fixed capital formation averaged 25.8 percent of GDP, comparable with other upper middle-income countries and fast-growing economies in the region. However, investment largely took place in non-tradable sectors like construction, retail and the financial sector. This stands in sharp contrast to other countries in the CEE region where manufacturing investments made up close to a quarter of all capital investment. The rise in consumption and investment was largely debt financed, implying a decline in the saving rate and leading to a significant increase in financial liabilities of house- holds, firms and the public sector. Rapid growth of domestic demand triggered an import surge, while performance of the export sector was modest. The contribution of exports to overall GDP growth was by far the lowest compared to Croatia’s peers in the CEE region and, besides Romania, Croatia was the only country where the share of exports in GDP did not rise in this period. This explains the negative contribution of net foreign demand to growth throughout the period and has resulted in a significant increase of the current account deficit and gross external debt. The global crisis pushed the economy into a prolonged recession and reversed the convergence process. Borrowing costs increased, capital inflows dried up, and the economy was pushed into a vicious circle of a sharp contraction in employment, private sector illiquidity and eroding business and consumer confidence leading to a sharp fall in investment and private consumption. A six-year long recession reduced output by 12 percent and investment by 33 percent, while unemployment doubled to 17.3 per- cent by 2014 and youth unemployment reached 50 percent. This implied a reversal of the convergence process, and by 2014 Croatia’s GDP per capita had dropped by 4 percentage points relative to the EU28 from its peak in 2008. In contrast, all other countries in the CEE region (except Slovenia) continued to converge towards the average EU28 level of income. 1 Refers to purchasing power standard, according to the Eurostat methodology. 2 In this note we define real convergence as β-convergence. This type of convergence occurs when lower-income countries grow faster than higher-income economies. We measure real convergence by comparing GDP per capita based on purchasing power parities for a single country with an average for the whole European Union. 3 Throughout this paper, we use CEE countries (Bulgaria, Czechia, Estonia, Hungary, Latvia, Lithuania, Poland, Romania, Slovenia and Slovakia) as comparators to Croatia. Macroeconomic stability, fiscal policy and taxation 6 Figure 1. GDP per capita, percent of EU28 GDP Figure 2. GDP growth and contributions to per capita, in PPS growth 100 15 90 SI CZ 10 CZ SI 80 SI SK 5 LT EE CZ 70 SK 0 LT EE HU PL LV % 60 HU RO -5 HU LV PL 50 SK RO -10 PL EE BG 40 LT -15 LV BG 30 BG -20 2010 2011 2002 2003 2004 2005 2006 2007 2008 2009 2012 2013 2014 2015 2016 2017 2018f 2019f 2020f 2021f RO 20 2006 2009 2012 2015 2001 2002 2003 2004 2005 2007 2008 2010 2011 2013 2014 2016 2017 Private consumption Residual item Change in inventories Croatia CEE Government consumption Source: Eurostat, WB staff calculations. Source: CBS, WB staff calculation In 2015, a more favorable external environment, along with the effects of EU accession in 2013, helped to reignite growth. In 2015, strong exports, as well as a pick-up of private consumption and investment led to 2.4 percent GDP growth, the first positive rate in six years. GDP growth was broad based, accelerating to 3.5 percent in 2016, before slowing down to 2.9 percent in 2017 and further to 2.6 percent in 2018. Exports performed strongly throughout this period, supported by record-high tourist seasons and a recovery of merchandise exports, as Croatian firms integrated in European value chains and increased their market shares after EU accession (see Ranilovic, 2017 and Section 4 for more infor- mation). At the same time, the rise in domestic demand was built on more solid foundations than before the crisis. Income tax cuts, low energy prices and the conversion of Swiss franc-denominated housing loans4 initially boosted private consumption, later further underpinned by favorable labor market devel- opments, while households continued to deleverage (Figure 3). Also, private investment recorded solid growth during the recovery period, accompanied by significant corporate deleveraging, even as govern- ment investment plunged due to the weak absorption of EU funds and fiscal consolidation (Figure 4). Higher employment, together with strong migration outflows, have resulted in a significant drop in unemployment. The economic recovery led to rising employment: the employment rate jumped from 42 percent in 2013 to 47 percent in 2018, with manufacturing, tourism and construction accounting for more than a half of the increase. Together with large negative migration flows these developments have led to a sharp drop in the unemployment rate, from 17.3 percent in 2014 to close to 8.4 percent in 2018. In addition, youth unemployment significantly decreased from 50 percent in 2013 to 26 percent in 2018, while the share of long-term unemployment has also fallen to 41 percent of total unemployment in 2018 from 63.7 percent at its peak in 2012. This has resulted in labor shortages in sectors that have witnessed a relatively strong recovery and relied heavily on domestic migration flows (e.g. tourism and 4 After the conversion of the Swiss franc denominated loans to loans in kuna or euro, monthly mortgage payment were sub- stantially reduced, freeing part of households’ disposable income for consumption and giving a boost to overall private spending. Macroeconomic stability, fiscal policy and taxation 7 construction). Thereby, labor supply has become a constraint on current growth but also for the econ- omy’s growth potential in the future (see Section 3). Figure 3. Private consumption financing Figure 4. Investment, real values 20 30 30 25 15 25 20 10 15 20 5 10 15 5 0 0 10 -5 -5 -10 -10 5 -15 2007 2010 2015 2005 2006 2008 2009 2011 2012 2013 2014 2016 2017 0 -20 2007 2010 2013 2016 2002 2003 2004 2005 2006 2008 2009 2011 2012 2014 2015 2017 2018 Real net disposable income, y-o-y Household consumption, y-o-y Private investment, % of GDP Household credits (without housing loans), y-o-y Public investment, % of GDP Share of household credists in net disposable income, Total Investment, growth rate - RHS y-o-y Private investment, growth rate - RHS Source: CNB, CBS, CPII, World bank staff calculation Note: Deflator for the public investment is estimated at the same value as for the total investment. Values for public and private investments for 2018 are estimates. Source: CBS, WB staff calculation On the supply side, during the entire period since 2001, the economy has undergone an – albeit modest – structural shift. The contribution of services to total value added (VA) rose from 65 percent in 2001 to below 71 percent of VA in 2018, while agriculture and industry declined from 6 and 29 per- cent of VA in 2001 to 3.6 and 25.8 percent, respectively. Industry, construction and trade have been the key sectoral drivers of gross value-added variations, with large contributions to growth prior to the crisis and a sharp contraction during 2009-2013. The recovery after 2015 appears broad-based across sectors, with a surge in export-oriented industry, and a recovery of trade and hotel and restaurant services, boosted by a robust tourism performance (Figure 5). This transformation however has been accompa- nied by only a modest increase in productivity growth, compared to pre-crisis times. Macroeconomic stability, fiscal policy and taxation 8 Figure 5. Gross value added, contributions to value added, sector composition, 2001-2018, per- centage points 6 4 2 0 -2 -4 -6 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 Construction Transport Agriculture Industry Hotels and restaurants Trade Financial intermediation Public sector Other GVA Note: Contributions for Hotels and restaurants and Public sector for 2017-18 are estimates. Source: CBS, WB staff calculations. Despite recent improvements, the recovery of the Croatian economy after the 2008 crisis started later and was weaker than in other CEE countries. The process of real convergence resumed in 2015, but at much slower pace than before the crisis, and it has not yet been sufficient to compensate for the divergence during the recession. As a result, in 2017, the last year for which the data are available, Croatia was the only country in the CEE region, besides Slovenia, whose relative GDP per capita had not yet reached the pre-crisis ratio. In comparison, Lithuania, which started from the same level in 2008 as Croatia (63 percent), in 2017 reached almost 80 percent of the average EU28 level of income and similar developments were recorded in most of the other countries in the region. Macroeconomic stability, fiscal policy and taxation 9 3 Potential Output At the current pace of potential growth, it would take decades for Croatia to reach current average in- come per capita in the EU28. Therefore, it is critical to raise the growth potential of the economy. How- ever, recent potential growth estimates are low, reflecting adverse demographic developments as well as lower investment. Moreover, numerous factors limit efficiency improvements in the economy, which is reflected in a very low total factor productivity (TFP) contribution to potential growth. Croatia’s potential GDP growth before the crisis was among the lowest among CEE peers, and with the smallest contribution of productivity.5 Despite relatively high growth rates in the 2002- 08 boom period, potential growth averaged 3.3 percent, compared to 4.7 percent in the CEE region, the lowest rate after Hungary (Figure 6). This largely stemmed from sluggish productivity growth, measured as TFP, which made the smallest contribution to potential growth in Croatia, at only 0.6 percentage points on average, compared to 2.7 percentage points in the CEE region. By contrast, the contribution of capital accumulation was broadly in line with other countries in the CEE region and the labor contri- bution was above the average of peer countries. Potential growth has remained low even in the recent recovery. During 2015-17, according to EC estimates,6 the growth rate of potential output was only 1.4 percent per year on average, the lowest among regional peers and 1.3 percentage points lower than the average for the CEE region (Figure 7). The labor contribution had turned negative (-0.2 percentage points) due to the unfavorable demographics and adverse migration trends, while the contribution of capital accumulation was still around the average of the CEE region. Although the TFP contribution improved compared to the pre-crisis period and be- came the leading factor of growth, it is still among the lowest when compared to the regional peers. 5 Potential output measures the level of output that fully utilizes available factors of production and is consistent with stable inflation. The output gap is the difference between current and potential output levels and helps distinguish cyclical and trend components of GDP growth. In the note, potential output estimates use the European Commission's production function ap- proach that factors in (i) the capital stock of the business sector; (ii) a measure of potential labor input based on the NAIRU estimate, the working age population and the trend labor force participation rate; and (iii) a measure of trend total factor productivity (TFP). 6 For the methodology, see http://ec.europa.eu/economy_finance/publications/economic_paper/2014/pdf/ecp535_en.pdf. Macroeconomic stability, fiscal policy and taxation 10 Figure 6. Potential Growth (Contribution to po- Figure 7. Potential Growth (Contribution to po- tential growth, percentage points), 2002-2008 tential growth, percentage points), 2015-2017 average average 7.0 4 6.5 6.0 5.5 3 5.0 4.5 4.0 2 3.5 3.0 1 2.5 2.0 1.5 0 1.0 0.5 0.0 -1 -0.5 -1.0 -1.5 -2 EU10 RO CZ LT LV HU HR BG EE PL SK SI EU10 LV LT RO CZ HU HR EE BG SK PL SI Labor Capital Productivity Potential growth Labor Capital Productivity Potential growth Source: EC, WB staff calculations. Labor is increasingly becoming a constraint for Croatia's growth potential.7 This reflects a shrinking labor force, due to adverse demographic trends, net emigration, as well as a still-low labor force participation rate. • Demographics: Croatia’s birth rate has declined from 9.2 to 8.9 over 2001-17. At the same time, the society is aging: the share of people aged over 65 has increased from 16.1 to 20.1 during the same period. This has led to an increase in the old-age dependency ratio (over 65-year olds to 15-64 years olds) from 39.8 to 43.7 percent. • Migration: Although official emigration data from the Croatian Bureau of Statistics (CBS) are lower due to the applied methodology,8 recent estimates suggest that close to 270,000 mostly young people left the country after accession to the EU (more than 67,000 annually on average) which is over 15 percent of the labor force (Figure 8). While roughly 115,000 people immigrated into Croatia during the same period, the negative net effect was still large. • Labor force participation: Furthermore, Croatia’s activity and employment rates are the second and the third lowest in the EU, respectively, and by far the lowest among CEE peers (Figure 9) 7 See the Policy Note Labor Market in Croatia for more information. 8 The statistical definition of international migrant used by CBS is based on the concept of usual residence. According to this concept, immigrants from abroad and emigrants to foreign countries are persons who have changed their country of usual residence for a period that is, or is expected to be, of at least one year. In practice, CBS uses the Ministry of Interior’s data on the number of persons who were absent for more than a year from their permanent residence for temporary stay outside the Republic of Croatia and have reported their absence to the Ministry of the Interior. Macroeconomic stability, fiscal policy and taxation 11 Figure 8. Net migration, thousands of persons Figure 9. Employment and activity rates, LFS, 2018 0 80 70 -10000 60 50 -20000 40 -30000 30 20 -40000 10 0 Poland Bulgaria Slovenia Lithuania Romania Slovakia Hungary Croatia Latvia Czechia EU Estonia -50000 -60000 2013 2014 2015 2016 Employment rate Activity rate Source: Dynamics and Determinants of Migration – The Case of Source: Eurostat Croatia and Experience of New EU Member States (Drazeno- vic, Kunovac and Pripuzic) Capital accumulation’s contribution to potential growth is unlikely to reach pre -crisis levels, despite the recent recovery of investment. During 2002-08, investment was one of the main drivers of GDP growth, with an average share in GDP of 26 percent (government investment reached 6 percent of GDP) and average annual growth above 10 percent. Concomitantly, rapid capital accumulation made the largest contribution to potential growth, also reflecting strong capital inflows from abroad. In the post-crisis period, while investments started to recover in 2015, the pace of capital accumulation is still much lower compared to the pre-crisis period. On the other hand, the composition of investment has been improving. Compared to 2002-08 -- when investments in large infrastructure projects, housing and retail dominated – the share of the tradables sector has increased (Figure 11). Going forward, the con- tribution of investment to potential growth will depend on Croatia’s attractiveness as an investment destination, as well as a recovery in public investment. Macroeconomic stability, fiscal policy and taxation 12 Figure 10. Investment share in the GDP and Figure 11. Change in investment composition, net FDI, percent of GDP 2015-2017 compared to 2002-2008, percentage points 30 8 6 25 4 20 2 0 15 -2 10 -4 -6 5 -8 0 -10 Real estate Energy Trade Other services ICT Construction Agriculture, forestry and fishing Manufacturing Banking Mining and quarrying Transportation and storage Public administration, education, health Accommodation and food service activities 2002-2008 2015-2018 Share of gross fixed capital investment, % of GDP Net FDI, % of GDP Source: CBS, CNB Source: CBS, WB staff calculations A greater contribution of TFP to potential growth will be crucial, as it is the only sustainable source of long-term growth. Croatia managed to improve its TFP growth in the last 15 years only marginally, and its rate remains among the lowest in the CEE region (Figure 12). Furthermore, the level of TFP is only 5 percent higher than in 2002, while the cumulative average increase in TFP in other CEE countries over the same period was around 30 percent (Figure 13), albeit with significant differ- ences among countries. Drivers of TFP are difficult to identify as TFP itself cannot be observed, so estimates of TFP should be interpreted with care. Nevertheless, the empirical literature points to quality of labor and capital inputs in the production process, innovative capacity of the domestic economy, as well as market institutions and regulation as important drivers of TFP (Figure 14)9. In many of these dimensions, there is large scope for improvement in Croatia: • Mobility of production factors. Firms, capital, labor and ideas face significant constraints on their mobility in Croatia. Low mobility of firms and capital, which impedes business dynamism, is related to business environment constraints like (1) entry-exit constraints, (2) level of state con- trol and product market regulation, (3) resolving insolvency and enforcing contracts, (4) access to credit and level of financial frictions and (5) dealing with construction permits and registering property. Also, product market regulation (PMR) indicators mark Croatia as one of the most regulated economies, with state control among the most stringent compared to regional peers. 9 See, for example, Balta (2013, 2014), Gabriel and Rosenblatt (2014) and Levenko, Oja and Staehr (2017). Macroeconomic stability, fiscal policy and taxation 13 • Resource allocation. The Croatian economy has a higher share of the services sector, and espe- cially low technology—and low-productivity—services, like tourism, distribution industries and construction, and a lower share of ICT and manufacturing compared to the CEE average (Fig- ure 15). Despite the recent recovery, Croatia's weak capacity to attract greenfield FDI to manu- facturing has persisted, resulting in the lowest greenfield investment within the CEE region (Fig- ure 16). • Education. Empirical research demonstrates that a higher share of skilled labor promotes inno- vation and can lead to higher TFP growth (Figure 25). Croatia has one of the lowest shares of workforce with tertiary education in CEE, and education outcomes are below the CEE average, as indicated by results in international tests (PISA, TIMSS). • R&D spending and innovation. Croatia has made no progress towards reaching its target for 2020, while most of its peers have not only progressed but have set more ambitious targets. Po- land, Slovakia, Hungary and Lithuania, which were trailing or on par with Croatia’s level of R&D spending in 2008, have now surpassed Croatia (Figure 16). Indirect innovation outcome indicators, such as scientific production, new trademarks, industrial designs and patents remain lackluster (Figures 18-22). • Digitalization. While Croatia has overall made good progress in recent years when it comes to digital performance, it still belongs to the low-performing cluster of EU countries. Croatian citi- zens are above-average internet users and enterprises are also keen to employ digital technolo- gies, but low connectivity remains a challenge, as rural broadband connectivity and fast broad- band coverage are limited, and affordability of broadband remains the lowest in Europe. Figure 12. Total factor productivity, average Figure 13. Total factor productivity, 2002=100 growth rate, percent 6 135 5 130 125 4 120 3 115 2 110 1 105 0 100 Lithuania Poland Bulgaria EU10 Romania Slovakia Slovenia Czech Republic Hungary Latvia Croatia Estonia 95 90 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2002 2003 2004 2005 2015-2018 2002-2008 EU10 Croatia Source: AMECO, WB staff calculation Macroeconomic stability, fiscal policy and taxation 14 Figure 14. Output development factors Source: van Ark, B. (2014), Total factor Productivity – Lessons from the Past and Directions for the Future, The Conference Board presentation Figure 15. Share of manufacturing and ICT in Figure 16. Investment capital pledged in green- GVA field projects by country, % of GDP 35 3.0 4.0 30 2.5 3.5 2.0 25 3.0 1.5 20 1.0 2.5 15 0.5 2.0 0.0 10 -0.5 1.5 5 -1.0 1.0 0 -1.5 0.5 Latvia Romania Poland Lithuania Bulgaria Czech Republic Hungary Slovenia Slovakia EU 10 Croatia Estonia 0.0 Poland Bulgaria EU 10 Lithuania Hungary Romania Czech Republic Slovenia Slovakia Latvia Croatia Estonia Share of manufacturing and ICT in the GVA, 2015-18 Change of share in 2015-18 compared to the 2002- 08 period - RHS 2014 2015 2016 Source: Eurostat, WB staff calculations. Source: FDI in Central, East and Southeast Europe: Recovery amid Stabilizing Economic Growth, WIIW, FDI markets, Euro- stat, WB staff calculations Macroeconomic stability, fiscal policy and taxation 15 Figure 17. R&D expenditure, 2017 Figure 18. Summary Innovation Index, 2017 2.5 2.0 1.5 1.0 0.5 0.0 EU28 HU RO CZ LT HR LV SK EE PL BG SI R&D expenditure in the business sector R&D expenditure in the public sector R&D total, 2008 Source: European Innovation Scoreboard Figure 19. Scientific Figure 20. PCT patent Figure 21. Trademark Figure 22. Design ap- publications among applications per bil- applications per bil- plications per billion the top 10% most lion GDP (in PPS), lion GDP (in PPS), GDP (in PPS), 2017 cited publications 2015 2017 worldwide (% of to- tal), 2015 12 4.0 18 7 3.5 16 6 10 3.0 14 5 8 12 2.5 10 4 6 2.0 8 3 1.5 4 6 2 1.0 4 2 1 0.5 2 0 0.0 0 0 Latvia Latvia Bulgaria Bulgaria Latvia Bulgaria Croatia Estonia Estonia Poland Hungary Estonia Croatia Poland Bulgaria Poland Slovenia Slovenia Hungary Slovenia Slovenia Latvia EU Slovakia Hungary Slovakia Romania Lithuania Lithuania Slovakia Estonia Czech Republic Lithuania Romania Romania EU EU Poland Czech Republic Czech Republic Slovakia Hungary Croatia EU Lithuania Romania Croatia Czech Republic Source: European Innovation Scoreboard Macroeconomic stability, fiscal policy and taxation 16 Figure 23. Digital Economy and Society Index Figure 24. DESI 2018 – relative performance by (DESI), 2018 dimension 70 Country with 60 highest score 50 40 30 20 Country 10 with lowest score 0 Connectivity Use of Internet Integration of Digital Human Capital Digital Public Services EU28 CZ LT LV RO HR HU EU 10 EE SI SK PL BG Services Technology Integration of Digital Technology Use of Internet Services Connectivity Human Capital HR EU 10 EU Digital Public Services Source: EC, WB staff calculation Figure 25. Human resources index and TFP Figure 26. Institutions and TFP growth growth 2.5 2.5 RO LV 2.0 RO LV 2.0 SK SK y = 1.4611x + 0.272 SE TFP average growth 1.5 SE 1.5 CZ TFP average growth DE CZ SI LT SI y = 0.6531x + 0.2325 UK LT DE 1.0 AT 1.0 BL FI AT DK BE FI FR DK PL FR UK PL 0.5 0.5 BG ES IT ES IT BG HU HU 0.0 0.0 EE EL HR EE HR EL -0.5 -0.5 0.0 0.2 0.4 0.6 0.8 0.0 0.5 1.0 1.5 Institution quality Human Resources Note: The graph shows partial regression plot, i.e. it illustrates Note: The graph shows partial regression plot, i.e. it illustrates the relationship between average Human resources index and the relationship between average of World Governance Indica- average real TFP growth between 2001-17 conditional on the tors in 2000-17 and average real TFP growth between 2001-17 initial level of real GDP per capita PPS in 2000. Human Re- conditional on the initial level of real GDP per capita PPS in sources is a composite index from European Innovation Score- 2000. Institution quality is an average score of WGI components board containing new doctorate graduates, population com- including Voice and Accountability, Political Stability, Govern- pleted tertiary education and lifelong learning. ment Effectiveness, Regulatory Quality, Rule of Law and Con- Source: European Scoreboard, Eurostat, WB staff calculation trol of Corruption. Source: WB Governance indicator, Eurostat, WB staff calcula- tion Macroeconomic stability, fiscal policy and taxation 17 Figure 27. Convergence scenarios, GDP per Figure 28. Projected GDP per capita, PPS, 2030 capita, PPS 50000 45000 Slovenia in Austria in 40000 45000 2028 2017 Germany in 35000 40000 2017 30000 Czech 35000 Slovenia in Republic in 25000 2017 2017 20000 30000 Slovakia in 2017 15000 25000 Poland in 2017 10000 20000 5000 15000 0 Lithuania Poland Slovenia Romania Latvia Bulgaria Czech Republic Slovakia Hungary Croatia Estonia 10000 2028 2030 2032 2014 2016 2018 2020 2022 2024 2026 2034 2036 2038 2040 2042 2044 2046 HRV @1.3% HRV @2.0% HRV @3.0% Source: Eurostat, WB staff calculation. Notes: Projection was based on average GDP potential growth rate in 2015-17 with constant population. Source: Eurostat, WB staff calculation. Unless it improves its growth potential, Croatia will need decades to reach the current level of living standards of the most developed EU countries or even some of its peers in CEE. For example, if the economy were to grow at the growth rate of 3 percent, as recorded in 2015-2017 period, it would take Croatia a full decade to reach a current living standard of Slovenia. This time almost triples if GDP growth slows down closer to the potential. (Figure 27). Raising the economy’s growth potential is critical for income convergence to proceed at a reasonable pace. To this end, incentives for higher labor force participation would need to be increased through reform of taxation and benefits. Also, and more broadly, reducing the outflow of workers could be achieved through state reform and better public service provision, as well as a better business envi- ronment. With regard to capital accumulation, raising public investment (without increasing overall spending—see below) could play an important role in catalyzing private investment. Lastly, a broad set of reforms to boost overall productivity and improve the business climate is needed. The reform agenda should be focused on the mobility and quality of production factors. A higher degree of factor mobility would improve capital allocation and reduce market frictions and could be achieved by improving the cumbersome business environment, reforming regulation, and increasing government effectiveness and quality of public services. At the same time, increasing the quality of human capital through educational reform, incentivizing R&D spending, developing innovation ecosystem and a higher level of digitaliza- tion would raise TFP through increases in the quality of factors of production. Macroeconomic stability, fiscal policy and taxation 18 4 Exports of Goods and Services Firm level data show that export-oriented firms tend to be the more productive, and have higher sales, wages, and utilization of capital10 than other firms. Hence, the export sector plays a crucial role in boosting productivity and incomes. Croatia’s export sector is small, but after EU accession there is sig- nificant potential for its development. Croatia has the smallest goods export sector among all CEE countries. During 2002-08, the average share of exports of goods was 16.2 percent of GDP, well below the CEE average of 39.5 per- cent. Therefore, exports had only a marginal impact on overall GDP growth. Also, Croatia’s presence on the international market did not increase, unlike its peers’, which used EU accession to better inte- grate into European and GVCs. The situation started to improve with EU accession in 2013, as Croatian firms used the opportunity to integrate into EU and global value chains. During the first years of its EU- membership, Croatia has increased its export market share by almost 20 percent (Figure 29). Figure 29. Export world market share, 2004=100 Figure 30. Exports of goods, percent of GDP 90 300 280 80 260 70 240 220 60 200 50 180 40 160 140 30 120 20 100 80 10 60 0 Lithuania Croatia Poland Slovenia Bulgaria Latvia Romania Hungary Czech Republic Slovakia EU 10 Estonia 2007 2011 2015 2004 2005 2006 2008 2009 2010 2012 2013 2014 2016 2017 Croatia Czech Republic Estonia Latvia Lithuania Poland Romania Slovenia 2015-2018 2002-2008 Slovakia Hungary Note: Bulgaria is omitted due to the deficiency of the data for Source: Eurostat, WB staff calculation the observed period. Source: Eurostat, WB staff calculation In 2015-2018, exports of goods became one of the main drivers of GDP growth. The average annual growth rate of goods exports in 2015-18 reached 7.8 percent and its contribution to GDP growth amounted to 1.5 percentage points. Furthermore, exports of goods recovered relatively quickly from the crisis and in 2018 was the only component of aggregate demand (except a minor increase in services exports and government consumption) that was on a significantly higher level —close to 24 percent of GDP compared to 15.6 percent of GDP in 2008 in volume terms—than in the pre-crisis period (Fig- ure 29). However, even these favorable developments compare poorly with regional peers, as the aver- age level of exports of goods in the CEE region in 2018 was 16.5 percentage points higher than in 2008. 10 Valdec, Zrnc (2015) Macroeconomic stability, fiscal policy and taxation 19 Also, while annual export growth rates in other countries in the region were lower than Croatia’s, their contribution to GDP growth was significantly higher, due to much larger export sectors (Figure 31). Figure 31. Exports of goods, growth rate and contributions to the GDP growth, percentage and percentage points 20 10 18 9 16 8 percentage points 14 7 12 6 10 5 % 8 4 6 3 4 2 2 1 0 0 Slovenia Poland Bulgaria Lithuania Romania Croatia Czechia Hungary Latvia Slovakia EU 10 Estonia Export growth rate, 2015-2018 Export growth rate, 2002-2008 Exports contribution to the GDP growth, 2015-20178 - RHS Exports contribution to the GDP growth, 2002-2008 - RHS Source: Eurostat, WB staff calculations Most of Croatia’s leading export sectors have increased their shares in global exports. During 2013-17, the majority have seen increased shares in global exports in sectors where global imports are increasing (Figure 32: “winners in growing sectors”, top-right quadrant). This suggests a remarkable change from 2008-12, when the majority of Croatia’s top exports were in declining sectors. However, for most of its main exporting products Croatia is the net importer, which implies high import-depend- ency of exports, and high re-export rate, and also potentially limits these sectors’ net contribution to growth. Macroeconomic stability, fiscal policy and taxation 20 Figure 32. Growth of national supply and international demand for products exported by Croa- tia, 2013-17 Source: ITC calculations based on UN COMTRADE and ITC statistics. The structure of Croatia’s exports has overall improved, albeit only to a limited extent: • Markets (Figure 33): The share of Italy, although it remains the largest destination for Croatia’s exports, has declined, largely on account of demand factors, namely the weak performance of the Italian economy, and developments in the Croatian shipbuilding industry. At the same time, ex- ports to other CEE countries have risen significantly, reflecting economic integration with neigh- boring countries after EU accession. • Value chains (Figure 34): While Croatia’s participation in GVCs remains low compared to peers, it has improved.11 One reason for Croatia’s level of GVC integration is the absence of an auto- motive sector; Croatia missed much of the foreign direct investment from Western Europe and Asia that entered the CEE region and spurred the development of the automobile manufacturing industry. Nonetheless, rising trade with CEE countries could point to increasing participation in GVC nodes in the periphery of the EU, while Croatia’s emerging success in several industrial machinery and electronic products, where sourcing products across borders (or longer distances) is common, reveals potential for entering new GVCs. 11 Perusko, Kovac, Josic (2018) Macroeconomic stability, fiscal policy and taxation 21 Figure 33. Exports of goods by markets Figure 34. Global Value Chain Participation in- dex, 2000 and 2014 100% 70% 90% 60% 80% 50% 70% 40% 60% 30% 50% 20% 40% 10% 30% 0% 20% Bulgaria Poland Hungary Slovenia Lithuania Czech Republic Slovakia Romania Latvia Croatia Estonia EU 10 10% 0% 2002-2008 2015-2017 Other Bosnia & Herzegovina CEE Italy Germany Austria GVC participation index 2014 Other EU 15 GVC participation index 2000 Source: UNCTAD, WB staff calculation Source: Croatia in Global Value Chains, Perusko, Kovac, Josic, CNB (2018), WB staff calculations • Technological complexity (Figure 35). The share of high-technology manufactures12 in Croatia’s export basket has been rising for the last four years and in 2017 reached almost 18 percent (com- pared to 12 percent in 2008). This is mainly due to the strong rise in export of pharmaceuticals, power generating machinery and turbines. At the same time, the share of medium-technology manufactures is falling and in 2017 had dropped to 30 percent (down from close to 40 percent in 2008), reflecting the collapse of Croatia’s shipbuilding industry that used to be one of the strong- est export sectors. Overall, the share of high- and medium-technology sectors combined re- mained broadly constant. Nonetheless, over the entire 2002-17 period, Croatia (together with the Baltic countries) is at the bottom, in the CEE, in terms of the share of high- and medium-tech- nology exports. 12High-technology manufactures include data processing and telecommunications equipment, television sets, transistors, turbines, power generating equipment, pharmaceuticals, aerospace, optical and instruments, cameras. Medium-technology manufactures include passenger vehicles and parts, commercial vehicles, motorcycles and parts, synthetic fibers, chemicals and paints, fertilizers, plastics, iron and steel, pipes and tubes, engines, motors, industrial machinery, pumps, ships, watches. Low-technology manufactures include textile fabrics, clothing, footwear, leather manufactures, travel goods pottery, simple metal structures, furniture, jewelry, toys, plastic products. Natural resource-based manufactures include prepared meats/fruits, beverages, wood products, vegetable oils, base metals (except steal), petroleum products, cement, gems, glass. Macroeconomic stability, fiscal policy and taxation 22 Figure 35. Technological complexity of exports 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Resource-based manufactures Low technology manufactures Medium technology manufactures High technology manufactures Source: UNCTAD, WB staff calculation Despite these improvements in the structure of exports, Croatia’s economic complexity re- mains low compared to most CEE peers and has been deteriorating. The Economic Complexity Index (ECI) measures the knowledge intensity of a country’s export looking at export diversification and export products’ ubiquity. For example, countries with high ECIs export many goods that are of low ubiquity and that are produced by highly diversified countries, indicating that these are diverse and sophisticated economies (Figure 36). In 2016 Croatia ranked 32nd out of 127 countries included in the analysis and has lower economic complexity compared to the average of the CEE region (Figure 36). Also, Croatia recorded the largest decline in economic complexity in the period between 2008 and 2016 (Figure 37). Figure 36. Economic complexity index rank, Figure 37. Change in Economic Complexity In- 2016 dex, 2016 vs 2008 1.8 0.30 1.6 0.25 1.4 0.20 1.2 1 0.15 0.8 0.10 0.6 0.05 0.4 0.2 0.00 0 -0.05 CZE HUN SVN SVK POL EU10 ROU EST HRV LTU LVA BGR -0.10 -0.15 -0.20 LVA CZE EST ROU LTU EU10 SVN SVK BGR HUN POL HRV Source: Atlas of economic complexity, WB staff calculations Source: Atlas of economic complexity, WB staff calculations Macroeconomic stability, fiscal policy and taxation 23 Croatia has a large services export sector, reflecting a strong tourism industry. The share of export of services in GDP in Croatia is the largest among CEE countries (Figure 38), and tourism ac- counts for around 70 percent of total exports of services. Exports of other services, including telecom- munication and computer services, construction and financial services and insurance, have risen over recent years and have reached close to 20 percent, while the share of transport and manufacturing ser- vices dropped. The strong performance of the tourism sector is reflected in the marked increase in the revealed comparative advantage (RCA) index for personal and recreational services (which includes tourism), while the RCA indices for other services sectors generally associated with higher value added, such as financial, telecom, or other business services, have remained stagnant. The total size and domi- nance of the tourism sector are one reason why the sophistication of Croatia’s services exports remains largely behind that of peers’. While a few countries have similarly high total services exports as a share of GDP, in no other country in the CEE region are they less diversified than Croatia. This weak diver- sification of Croatia’s services export is also reflected in the lowest share of ICT, transport and other business services in the CEE region (Figure 39). Figure 38. Exports of services, percent of GDP Figure 39. Exports of services, by service group, 2015-2017 30 100% 25 80% 20 60% 15 40% 10 20% 5 0% Bulgaria Croatia EU 10 Latvia Romania Estonia Hungary Poland Slovakia Czechia Lithuania Slovenia 0 Poland Slovakia Estonia Lithuania Slovenia Hungary Latvia Bulgaria Croatia Czechia Romania EU 10 Other services Financial services Constuction ICT Transport Manufacturing and maintenance 2015-2018 2002-2008 Travel Source: Eurostat, WB staff calculations Exports of tourism services recently recorded a strong expansion and have become one of the most important drivers of economic recovery in 2015-18. Average annual growth of travel services exports in 2015-18 amounted to 8.1 percent (in nominal terms) with a sharp rise in tourist arrivals and overnight stays. Revenues from tourism reached almost 20 percent of GDP, and the contribution of tourism to GDP, taking also indirect income effects into account, is estimated at 25 percent13 and on the rise. This makes Croatia the most tourism-dependent economy in the EU (Figure 40). Croatia’s attrac- tiveness as a tourist destination reflects large investments in accommodation and improvement in the overall quality of service, but also external factors, such as geopolitical tension in some of its competi- tors on the Mediterranean market (Figure 41). 13 Travel and Tourism Economic Impact 2018, Croatia, World Travel and Tourism Council Macroeconomic stability, fiscal policy and taxation 24 Figure 40. Tourism revenues and employment, Figure 41. Foreign tourist arrivals, index, 2017 2010=100 25 250 20 200 15 150 10 100 5 50 0 0 2014 2015 2016 2017 2007 2008 2009 2010 2011 2012 2013 Croatia Cyprus Malta Greece Spain Italy Tourism revenues - % of GDP Tourism emplyment - % of total employment Croatia Mediterranean Source: Eurostat, WB staff calculations Source: Eurostat, WB staff calculations However, Croatia’s tourism has limited spill-overs on the whole economy. It is highly concen- trated in the coastal areas and the summer months, resulting in the strongest seasonal profile in the EU.14 High seasonality of tourism leads to high seasonality of employment, resulting in a sizable proportion of temporary workers and, recently, labor shortages. Seasonality also contributes to the high reliance of Croatian tourism on imports, as during the short summer season the surge in demand surpasses local production capacities. As a result, exports of travel services are one of the main drivers of imports in Croatia,15 which reduces the positive effects of tourism on the overall economy. Mor eover, Croatia’s tourism sector purchases fewer inputs from other sectors (backward linkages), compared to other coun- tries (Figure 42), which suggests limited integration with the local economy.16 Tourism is less innovative than knowledge-intensive services, suggesting that the effects on the productivity of other sectors in the economy is limited. This could, to some extent, also explain the low total factor productivity contribu- tion to growth in Croatia. Finally, reliance on one sector, the performance of which strongly depends partly on external factors, exposes the entire economy to considerable risk. 14 Orsini, Vukasin (2018) 15 Orsini (2017) 16 See Policy Note Growth, Competitiveness and Innovation , Box 3. Croatia’s intensive tourism and i ts linkages. Macroeconomic stability, fiscal policy and taxation 25 Figure 42. Indirect forward and backward linkages in export value added as a share of total ex- ports, accommodation and food services, 2014 Source: Staff calculation based on the World Input-Output Database Removing constraints to exports development should be one of the main policy priorities. Ex- ports of goods were one of the main drivers of the Croatian economy in the 2015-2018 recovery as firms further integrated into EU value chains following EU accession, and exports became more diversified, complex, and gained market share. However, the export base remains the smallest among peers (requir- ing high exports growth rates for pulling the economy), while diversification, technological complexity of exports and participation in GVCs is still weak. Therefore, one of the policy priorities should be reducing excessive product market regulation, business environment bottlenecks and barriers to invest- ment. This could spur the expansion and creation of export-oriented firms, which tend to be more pro- ductive. At the same time, measures to increase linkages of exports of tourist services—with revenues estimated around 25 percent of GDP and record-high growth rates in 2015-2018—with the rest of the economy should be pursued. Macroeconomic stability, fiscal policy and taxation 26 5 Public finances During the crisis and long recession, public debt rose significantly, raising risks and limiting fiscal space. At the same time, the public sector is relatively large, end there is scope to improve the reve- nue and expenditure composition, while reducing the overall tax burden. 5.1 General government deficit and debt In the pre-crisis period (2002-08), fiscal policy was predominantly expansionary and frequently pro-cyclical (Figure 43). Even though the output gap was positive, expansionary fiscal policy amplified the boom and exacerbated internal and external imbalances. During the crisis, the prolonged and deep contraction of output resulted in large fiscal imbal- ances. Nominal revenues dropped while expenditures remained stable as a significant rise in interest payments and social spending (including unemployment benefits) was offset by a sharp cut in public investment. As a result, the general government deficit deteriorated to 6 percent of GDP on average in 2009-2014 (Figure 45). However, the large structural deficits as calculated by the European Commission (EC) over the 2010-201417 period suggest that the root causes of high deficits were of a structural rather than a cyclical nature. As high deficits accumulated, public debt more than doubled by the end of the recession, from 39 percent of GDP at end-2008 to a peak of 84 percent at end-2014 (Figure 44). This increase reflects not only the accumulation of high primary deficits, but also the positive difference between the nominal average interest rate that Croatia was paying on its debt and GDP growth (Figure 46), the activation of contingent liabilities as government took over some of the debt of public enterprises, as well as valuation changes due to changes in the exchange rate.18 17 Structural deficit estimates for earlier years are not available. 18 Croatia’s CDS increased from 124 (at end-August 2008) to above 300 basis points in 2013. Macroeconomic stability, fiscal policy and taxation 27 Figure 43. Cyclicality of fiscal policy, 2002-2018 Figure 44. Fiscal performance, percent of GDP, ESA 2010 5 90 84.083.7 1 80.4 80.5 Change of cyclical primary balance, percentage points Restrictive procyclical 77.8 4 Restrictive 80 74.6 0 2012 countercyclical 69.5 3 63.9 -1 70 57.3 -2 2 60 2016 2015 2017 48.3 -3 % of GDP % of GDP 1 2014 2005 50 2008 40.341.2 2009 38.1 38.737.339.0 -4 36.6 0 40 2013 2010 -5 2007 -1 2004 2006 30 -6 2018 2003 -2 2011 20 -7 2002 10 -3 -8 Expansionary Expansionary countercyclical procyclical 0 -9 -4 2006 2013 2002 2003 2004 2005 2007 2008 2009 2010 2011 2012 2014 2015 2016 2017 2018 -5 Domestic debt Foreign debt -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 Total public debt Fiscal balance - RHS Output gap, % Structural balance - RHS Note: Cyclicality is determined as a change of cyclical primary Source: Eurostat, CNB, EC, WB staff calculation balance related to the output gap in the given year based on the potential GDP. Source: EC, WB staff calculation Figure 45. Decomposition of the change of general government debt, percentage points of GDP 13 11 9 percentage points 7 5 3 1 -1 -3 -5 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 Difference between interest rate and GDP growth Primary balance Other Change of public debt Source: CNB, CBS, Eurostat, WB staff calculation Fiscal consolidation began in 2014, under the framework of the EU’s Excessive Deficit Proce- dure (EDP).19 Positive cyclical effects as the economy began to recover, together with tax policy 19Croatia remained under the EDP until 2017, after the nominal headline deficit fell to 0.9 percent of GDP (in 2016), and public debt was put on a downward path. In addition, the EC estimated that Croatia had improved its structural deficit by 3 percent of GDP in the period 2014-2016. Macroeconomic stability, fiscal policy and taxation 28 changes, resulted in an increase of total revenues from 43 percent of GDP in 2014, the last year of recession, to 46.6 percent in 2018. At the same time, expenditures remained stable in kuna terms over this period and declined from 48 percent of GDP in 2014 to 46.4 percent in 2018. This led to a reduction of the general government deficit from 5.1 percent in 2014, to 3.4 percent in 2015, 0.9 percent in 2016, finally reaching a surplus of 0.8 percent and 0.2 percent of GDP in 2017 and 2018, respectively. Fiscal consolidation reversed the unsustainable trajectory of public debt, but fiscal risks per- sist. As fiscal imbalances narrowed, and growth accelerated, public debt was reduced from its peak of 84 percent of GDP in 2014 to 74.6 percent at end-2018 (Figure 46). However, fiscal vulnerabilities re- main. Croatia’s public debt is the highest within the CEE region and almost double the group’s average, closer to the EU1520 high-income countries which are on a much higher development level (Figure 47). This renders Croatia highly vulnerable to a possible worsening of external borrowing conditions or in- terest rate shocks. In addition, as 76 percent of public debt is denominated in foreign currency, exchange rate risk is also present. Figure 46. Croatia, CEE and EU15: General Figure 47. Croatia, CEE and EU15: Interest Pay- Government Debt, Percent of GDP ments, Percent of GDP 100 4.0 89.0 90 84.0 3.4 81.6 3.5 80 74.6 68.8 3.0 70 2.4 2.5 2.3 2.3 60 2.1 46.1 48.3 1.9 50 2.0 40.8 1.6 1.5 40 33.9 1.3 1.5 30 1.0 20 10 0.5 0 0.0 EU15 CEE HR EU15 CEE HR 2009 2014 2018 2009 2014 2018 Source: Eurostat, WB staff calculations. Debt sustainability analysis (DSA) reveals that public finances are sensitive to various shocks, especially a growth shock and an exchange rate shock. Under the baseline scenario, public debt would continue its downward trajectory, falling below 60 percent of GDP by 2023. Downside risks to the baseline scenario are associated with (i) worsening of financing conditions reflected in an interest rate shock and EUR depreciation, (ii) a growth shock related especially to a deterioration in the external environment (including weaker EU growth), (iii) activation of contingent liabilities. The public debt trajectory is most sensitive to the growth shock, where a slowdown of the economy would result in an explosive debt trajectory. (Figure 48 and Table 1). While in the short run debt is firmly on a downward path, a long run perspective is needed as the external environment is expected to become less supportive 20EU15 includes Belgium, Denmark, Germany, Ireland, Greece, Spain, France, Italy, Luxembourg, Netherlands, Austria, Portugal, Finland, Sweden and United Kingdom. Macroeconomic stability, fiscal policy and taxation 29 and contingent liabilities in firms, and in the health system will eventually need to be addressed. There- fore, rebuilding fiscal space that could serve as a cushion in future downturns should be a priority. Figure 48. Scenario analysis – Croatia’s Table 1. Scenario analysis – Croatia’s Public Debt Public Debt (percent of GDP) (percent of GDP) 2018 2019 2020 2021 2022 2023 Baseline 74.6 70.6 67.3 64.2 61.2 58.3 Higher i-rate 74.6 72.5 71.0 69.6 68.3 67.0 GDP growth shock 74.6 74.4 76.1 79.6 84.6 91.3 20 % Depreciation 74.6 82.0 78.5 75.3 72.2 69.2 Combined i- 74.6 82.1 83.2 86.0 90.3 96.2 rate/growth/depreciation Contingent liabilities 74.6 72.2 68.8 65.7 62.7 59.8 Note: See Appendix II for details on shock design. Source: WB staff calculation These risks and the currently positive output gap suggest that now is a good opportunity to accelerate debt reduction. In this context, the authorities’ current plans of a mildly expansionary fiscal policy (deficits of 0.3 and a small surplus of 0.2 percent of GDP are planned for 2019 and 2020, respec- tively) should be reconsidered. 5.2 General government revenues Croatia has the highest tax burden on the economy among its peers. The share of general govern- ment revenues in GDP in Croatia is close to the highly developed euro area countries, higher than the EU average and the highest in the CEE group (Table 2 and Figure 49). Taxation levels are 5 percentage points of GDP higher in Croatia than its peers in the EU10. Moreover, Croatia stands out as having higher a share of revenues in GDP when compared to its income level (Figure 50). For example, the EU has on average a lower share of revenues but 40 percent higher income per capita than Croatia, while Germany, with the same revenues-to-GDP ratio, has double the living standards compared to Croatia. Table 2. General Government Revenues by Economic Classification, Percent of GDP, 2018 Euro area Croatia EU 10 Bulgaria Czechia Hungary Poland Romania Slovenia Slovakia Total Revenues 46.3 46.6 39.0 36.8 41.7 44.2 41.2 32.0 43.1 39.9 Direct taxes 13.7 7.6 7.6 6.6 8.6 7.3 8.4 6.1 8.0 8.9 Indirect taxes 13.0 20.1 13.5 14.8 12.4 18.3 14.1 10.4 10.9 14.1 Sales 3.2 4.2 3.0 2.8 3.3 3.5 2.3 2.2 3.9 4.2 Social contributions 15.2 12.0 12.6 8.8 15.7 12.2 14.1 11.4 14.8 14.8 Other current revenue 1.2 2.7 2.3 3.8 1.7 2.9 2.3 1.9 5.5 -2.1 Source: Eurostat, WB staff calculation. Macroeconomic stability, fiscal policy and taxation 30 Figure 49. Total general government revenues, Figure 50. General government revenues and by component, % of GDP, 2018 GDP per capita PPS (EU28=100) 50 55 45 FR FI DK General governement revenues, % of GDP 40 BE 50 SE 35 AT 30 EL IT EA19 HR 25 DE 45 PT 20 HU EU28 NL SI 15 PL CZ SK ES 10 40 CY UK 5 LV EE BG MT 0 Slovenia Poland Latvia Bulgaria Lithuania Euro area Hungary Croatia Czechia Slovakia Romania EU 10 EU Estonia 35 LT RO 30 40 60 80 100 120 140 Indirect taxes Direct taxes Social contributions Other GDP per capita PPS (EU28=100) Source: Eurostat, WB staff calculation Note: Luxembourg and Ireland are excluded as outliers. Source: Eurostat, WB staff calculation Croatia’s tax system differs from the patterns observed in other EU countries, largely because it is skewed toward indirect taxes. • Consumption taxes are high, yielding almost 20 percent of GDP in revenue, among the highest in the EU. Two-thirds of this is contributed by VAT revenue which at 13 percent of GDP has the highest share in the EU and, at 25 percent, is set at one of the highest general rates in the EU. While indirect taxes are economically more efficient than direct taxes, high consump- tion taxes have significant regressive effects, thereby counteracting redistribution through trans- fers. For Croatia, this implies that the Gini coefficient before and after taxes and transfers is nearly the same. Although the overall fiscal effort, including health and education spending, is inequality reducing, the overall redistributive effect is smaller than in other countries.21 • Tax burden on labor in Croatia is slightly below average for the CEE region, but progres- sivity is higher (Figure 51). The top personal income tax rate is 36 percent, the highest, after Slovenia, in the CEE region (Table 3.). Together with the surcharges and relatively high social contribution rates, this implies a relatively high tax wedge on higher incomes.22 Despite recent tax reforms, the tax wedge for top earners remains relatively high compared to peers (Figure 51.). 21Inchauste and Rubil (2017). 22The tax wedge is defined as the share of the total tax burden, including personal income tax and surtax as well as social contributions, in overall labor cost. A higher tax wedge tends to disincentivize employment, and can reduce a country’s com- petitiveness and attractiveness as an investment destination. Macroeconomic stability, fiscal policy and taxation 31 Table 3. VAT, PIT and CIT rates, percent, 2018 Top personal income Top corporate income VAT standard rate tax rates tax rates Bulgaria 20 10 10 Czech Republic 21 15 19 Estonia 20 20 20 Croatia 25 42 18 Latvia 21 31 20 Lithuania 21 15 15 Hungary 27 15 11 Poland 23 32 19 Romania 19 10 16 Slovenia 22 50 19 Slovakia 20 25 21 EU 10 21.4 22.3 17.0 EU 21.5 39.0 21.9 EA 20.8 42.6 24.1 Note: In top PIT and CIT tax rates are surcharges included. Source: EC Figure 51. Tax wedge, Croatia and EU 10, percent of total labor cost, 2017 60 60 55 55 50 50 45 45 40 40 % 35 % 35 30 30 25 25 20 20 EU10 HU LV LT CZ RO HR EE BG SI SK PL HU EU10 CZ HR LV EE SI SK PL 67% 100% 167% 300% 500% 500% (2019) Note: Tax wedges are evaluated at 67, 100, 167, 300 and 500 percent of average wage. Source: EC, CNB, OECD Taxing Wages 2018 • Taxation of capital in Croatia is similar as in regional peers, but there is a high burden on firms in the form of parafiscal charges (Figure 52). Over the last ten years, Croatia has reduced its effective tax rate, in line with regional trends. Today, both standard CIT rate (currently at 18 per- cent) and the effective marginal tax rate (slightly below 15 percent) are in similar to those found Macroeconomic stability, fiscal policy and taxation 32 in comparable countries in the CEE region(Figure 53).23 However, although in recent years there have been efforts for its reduction, the burden of parafiscal charges is still considerable, with 547 fees on the private economy, according to the latest Register of Non-Tax Revenues published by the Ministry of Economy.24 Figure 52. Taxes on income and wealth, per- Figure 53. Effective marginal average tax rate, cent of GDP, 2018 non-financial sector, percent, 2017 14 25 4 12 2 20 10 0 15 -2 8 10 -4 6 -6 5 4 -8 2 0 -10 Romania Hungary Poland Slovenia Bulgaria Lithuania Slovakia Czechia Latvia Croatia EA Estonia EU 10 0 Poland Slovakia Estonia Slovenia Croatia Lithuania Euro area Czechia Latvia Hungary Bulgaria Romania EU 10 Effective marginal average tax rate Difference in EMR from 2008 to 2018 - RHS Source: Eurostat, WB staff calculation Source: EC, WB staff calculation 5.3 General government expenditures Croatia's expenditure-to-GDP ratio is well above the level prevailing in most of its CEE peers. 25 In 2018, general government expenditure stood at 46.4 percent of GDP, 7 percentage points higher than in peer countries. This mainly reflects non-productive spending, while productive spending is at a com- parable level with the EU 10.26 Moreover, the fiscal adjustment required during the global financial crisis fell to a large extent on investment cuts, leading to one of the lowest levels of investment spending among CEE countries. 23 The effective marginal tax rate on firms is used, which takes a hypothetical investment and calculates the difference be- tween pre-tax return (cost of the capital) and post-tax real return on investment (EC, Taxation Trends in the European Union (2018), Annex B). 24 According to the Ministry of economy, total burden of fees adds up to 2.8 percent of GDP in 2016 and are shared between firms and households at 1.6 percent and 1.1 percent of GDP, respectively. 25 However, the scope of general government differs within the EU. Croatia has a lower ratio of liabilities of government cor- porations that are classified outside general government at 10.4 percent of GDP than the average for the CEE region at 18.5 percent of GDP in 2016. 26 Productive government spending is defined as spending that increases productive capacity of the economy (e.g. improves the quality and/or quality of human and physical capital) and is usually estimated as the sum of government expenditure on health, education and capital investment. Macroeconomic stability, fiscal policy and taxation 33 Figure 54. Productive and non-productive Figure 55. Expenditure-to-GDP ratio, 2002-2018 spending, Croatia and EU 10, 2017 45 55 40 50 35 30 in % of GDP in % of GDP 45 25 20 40 15 10 35 5 0 30 Productive Productive Non-productive Non-productive 2013 2014 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2015 2016 2017 2018 EU 10 Croatia EU 10 Croatia Health Education Croatia EU 10 Source: Eurostat, WB staff calculation Source: Eurostat, WB staff calculation Table 4. General government expenditures by economic classification EU Croatia Bulgaria Czechia Estonia Latvia Lithuania Hungary Poland Romania Slovenia Slovakia in % of GDP 10 2008 2015 2018 2018 2018 2018 2018 2018 2018 2018 2018 2018 2018 2018 Total 45.3 48.3 46.4 39.4 34.8 40.8 39.5 38.5 34.0 46.5 41.5 35.0 42.4 40.6 expenditure Total current 38.0 43.4 41.1 34.5 31.6 36.0 33.6 32.9 30.6 38.5 36.3 31.0 38.2 36.6 expenditure Current transfers 17.8 20.5 18.9 17.4 16.3 19.3 15.9 16.1 15.5 18.4 19.2 13.7 19.0 20.6 Social payments 14.0 17.0 15.6 14.3 13.2 14.9 13.5 11.5 13.5 13.5 16.8 11.7 16.5 18.1 Subsidies 2.5 1.4 1.2 0.5 0.9 1.7 0.3 0.2 0.2 0.8 0.0 0.3 0.1 0.1 payable Other current 1.3 2.1 2.1 2.6 2.2 2.7 2.1 4.4 1.8 4.1 2.4 1.7 2.4 2.4 expenditure Interest payable 2.0 3.5 2.3 1.2 0.7 0.8 0.0 0.7 0.9 2.5 1.4 1.2 2.0 1.3 Compensation of 11.1 11.4 11.7 10.2 9.5 9.8 11.3 10.2 9.8 10.5 10.1 11.0 10.9 9.3 employees Intermediate 7.1 8.0 8.2 5.7 5.1 6.1 6.4 5.9 4.4 7.1 5.6 5.1 6.3 5.4 consumption Total capital 7.3 4.9 5.3 4.8 3.2 4.8 5.9 5.6 3.4 8.0 5.2 4.0 4.2 4.0 expenditure Investment 5.9 3.2 3.5 4.1 2.9 4.1 5.5 5.4 3.2 5.8 4.7 2.6 3.6 3.6 Source: Eurostat Croatia stands out for high spending on general public services, health and social protection compared to its regional peers (Tables 4 and 5). Spending on general administration is 4 percentage points higher in Croatia than in EU10 countries. This partly reflects high interest expenses, which result from sizable public debt and also relatively high borrowing costs. Nevertheless, adjusted for the cost of debt servicing, spending on general administration is still considerably higher than in CEE countries. Health spending is one percentage point higher in Croatia than in its peers, predominantly due to higher compensation of employees, although this might partly reflect different insurance coverage. Spending on social protection is also one percentage point higher than in EU10 peers as a result of high social benefits, which can be in turn linked to higher pension expenditures. Available data suggests that this Macroeconomic stability, fiscal policy and taxation 34 can be largely explained by War veterans’ pensions, a type of spending that is not present in other coun- tries. In most other categories, Croatia’s expenditure-to-GDP ratio is close to that of its peers. Spending on economic affairs and spending on recreation, culture and religion are respectively 0.7 and 0.3 percentage points higher than in the EU10. Croatia also has somewhat higher level of spending on public order and safety and housing and community, but lower spending on defense, while spending on education, environmental protection is at a comparable level to the EU10. Table 5. General government expenditures by Functional Classification Croatia EU 10 Bulgaria Czechia Estonia Latvia Lithuania Hungary Poland Romania Slovenia Slovakia in % of GDP 2008 2017 2017 2017 2017 2017 2017 2017 2017 2017 2017 2017 2017 Total general 45.3 45.0 38.9 35.1 39.0 39.3 37.8 33.1 46.9 41.1 33.7 43.2 40.2 government expenditure General public services 7.7 7.6 4.7 3.2 3.9 3.9 4.1 3.5 8.0 4.4 4.2 5.9 5.6 Defence 1.6 1.1 1.4 1.1 0.8 2.0 1.7 1.7 1.0 1.7 1.8 0.9 1.0 Public order and safety 2.2 2.2 2.0 2.5 1.8 1.9 2.3 1.5 2.4 2.1 2.0 1.6 2.1 Economic affairs 7.0 5.4 4.7 4.0 5.7 4.3 5.5 2.8 7.1 4.7 4.4 4.3 4.1 Environmental protection 0.4 0.6 0.6 0.7 0.8 0.7 0.6 0.4 0.4 0.4 0.5 0.5 0.7 Housing and community 1.1 0.9 0.7 1.6 0.6 0.4 1.1 0.4 0.8 0.6 0.9 0.5 0.5 amenities Health 6.2 6.3 5.4 4.9 7.5 5.0 3.5 5.7 4.8 4.7 4.3 6.6 7.1 Recreation, culture and 1.00 1.8 1.5 1.0 1.3 2.1 1.7 1.1 3.5 1.2 1.0 1.4 0.8 religion Education 4.8 4.7 4.7 3.6 4.6 5.8 5.8 4.9 5.1 4.9 2.8 5.4 3.8 Social protection 13.5 14.3 13.3 12.5 12.0 13.0 11.7 11.2 14.0 16.4 11.7 16.2 14.5 Source: Eurostat Continued tight fiscal policy is needed to reduce fiscal risks and rebuild fiscal space, and there is scope to improve the composition of revenue and spending. Better targeting of social assistance could help raise labor force participation while reducing inequality and poverty. At the same time, sav- ings in administration (see below) would increase the scope for productive spending. Macroeconomic stability, fiscal policy and taxation 35 6 Public Sector Performance and Efficiency While Croatia spends more than its peers, performance of public sector is below average, implying that public spending is also inefficient. Boosting public sector efficiency could generate significant sav- ings, while improving the business environment. An analysis of public sector performance and efficiency in Croatia following the approach proposed by Afonso et al. (2005) indicates that Croatia is ranking low on all dimensions, compared to EU28 and CEE peers.27 While the results need to be interpreted with caution as the analysis does not separate the effects of public sector spending from other potential influences, and there are issues of data comparability, and whether cost and benefits of governmental activities are being properly meas- ured, the results are nonetheless suggestive. • Overall, Croatia is the worst performer among CEE peers. All CEE countries except Czechia score below overall sample mean. Looking at all countries in the sample, Croatia is at the bottom end of the list together with Greece, Italy and Latvia. • Looking at sub-indicators, Croatia is the worst performer in administration among all countries included in the sample. This reflects poor administration performance including cor- ruption, red tape, judiciary efficiency and the size of shadow economy. • Croatia also scores below the CEE average in most other “opportunity” indica tors. This includes weak performance in education and health, but is in part counterbalanced by above-CEE average performance in public infrastructure. • Regarding Musgravian indicators, Croatia scores poorly in stability and economic perfor- mance. However, on the distribution indicator, Croatia is slightly above the average CEE per- formance, though the variation between countries is not large. 27Public sector performance (PSP) is assessed by constructing PSP composite indicators which are based on socio-economic indicators that are assumed to be the output of public policies. Indicators are divided into two groups. The first group includes outcomes of public policies aimed at providing various and accessible public services for individuals. Afonso et al. describe these as “opportunity indicators”, referring to the role of the government as a promoter of equal opportunities in the market place. Opportunity indicators are divided into four sub-groups which reflect government performance in administration, edu- cation, health and infrastructure. The second group include indicators linked to the standard Musgravian functions of the gov- ernment - allocation, distribution and stabilization. Macroeconomic stability, fiscal policy and taxation 36 Table 6. Public Sector Performance – 2017 Administra- Public Economic Overall Country Education Health Distribution Stability tive infrastructure Performance Performance Poland 0.77 0.99 0.87 0.85 1.01 3.50 1.48 1.35 Czech Republic 0.84 0.97 1.11 0.93 1.08 1.00 1.18 1.01 Slovakia 0.75 0.87 0.80 0.85 1.08 1.32 1.17 0.98 Slovenia 0.78 1.01 1.37 0.92 1.09 0.63 0.80 0.94 Estonia 1.01 1.03 1.16 1.08 0.98 0.40 0.79 0.92 Romania 0.73 0.84 0.67 0.67 0.95 0.91 1.29 0.87 Hungary 0.77 0.94 0.83 0.91 1.03 0.71 0.85 0.86 Bulgaria 0.75 0.89 0.71 0.78 0.85 0.98 0.98 0.85 Lithuania 0.83 0.94 0.83 0.99 0.89 0.50 0.87 0.84 Latvia 0.74 0.99 0.86 0.86 0.93 0.21 0.48 0.73 Croatia 0.58 0.92 0.88 0.95 1.00 0.36 0.35 0.72 Average EU10 0.80 0.94 0.93 0.89 0.99 0.74 0.93 0.89 Source: WB staff calculation Over time, Croatia’s performance has regressed in relative terms (Figure 56). When comparing the overall PSP indicator for 2017 with 2007, Croatia has declined over the last ten years relative to other countries and the worsening of its performance was one of the largest among CEE countries. However, it is important to emphasize that change over time is measured relative to other countries, which means that a given country could achieve a higher score because of the improvement in certain areas or because other countries in the sample achieve weaker results. Figure 56. PSP indicators for 2007 and 2017. Performance 1.4 PL 1.2 Performance -2017 CZ 1 SK SI EE RO HU BG LT .8 LV HR .6 .6 .8 1 1.2 1.4 Performance - 2007 Source: W orld Bank Source: WB staff calculation Macroeconomic stability, fiscal policy and taxation 37 Relating general government expenditures to public sector performance indicators gives an indication of public sector efficiency. This implies that government spending (by functional classifi- cation) is considered as input for the relevant PSP indicators.28 The results largely mirror those of the PSP assessment (Table 7), both in overall performance and along sub-indicators, though the gap be- tween Croatia and the EU10 average is generally larger. One exception is distribution, where Croatia scores slightly above the EU10 in terms of performance but somewhat below average the EU10 average in terms of efficiency. Table 7. Public Sector Efficiency Indicators, 2017 Administra- Public Economic Overall Country Education Health Distribution Stability tion infrastructure Performance Performance Poland 0.74 0.97 1.16 0.68 1.01 3.56 1.55 1.38 Romania 0.77 1.27 1.06 0.50 1.24 1.08 1.57 1.07 Slovakia 0.63 1.14 0.72 0.82 1.13 1.40 1.27 1.02 Bulgaria 0.73 1.25 0.96 0.68 1.08 1.16 1.18 1.01 Czech Republic 0.71 1.02 0.93 0.76 1.23 1.05 1.25 0.99 Estonia 0.94 0.85 1.41 0.72 1.22 0.44 0.88 0.92 Lithuania 0.86 0.84 0.86 0.90 1.06 0.58 1.04 0.88 Slovenia 0.78 0.85 1.28 0.79 0.97 0.58 0.74 0.86 Hungary 0.61 0.97 1.04 0.81 1.03 0.64 0.78 0.84 Latvia 0.61 0.87 1.36 0.69 1.25 0.23 0.55 0.80 Croatia 0.47 0.97 0.86 0.90 1.05 0.34 0.33 0.70 Average EU10 0.74 1.00 1.08 0.74 1.12 1.07 1.08 0.98 Source: WB staff calculation To complement the efficiency analysis, we also apply Data Envelopment Analysis (DEA) using total spending-to-GDP ratios as an input and the overall PSP score as an output. DEA can be conducted for both input and output-oriented analyses.29 Figure 57 shows the theoretical production possibility frontier, which illustrates how far the distance is between the most efficient countries in the sample and countries that are below efficiency frontier. Croatia is the farthest from the best performers. 28 Administrative performance is weighted by government collective consumption; education and health performance are by spending in education and health, respectively; while quality of infrastructure is weighted by general government investment expenditures. As for the Musgravian functions, distribution performance is weighted by spending on social transfers other than social transfers in kind, while stability and economic performance are related to total general government spending. 29 Input oriented specification shows how much input quantities can be proportionally reduced without changing the output quantities. Conversely, output oriented specification tries to assess how much output can be proportionally increased without changing the input quantities used. The two measures provide the same results under constant returns to scale but give differ- ent values under variable returns to scale. For detailed analytical description see Afonso et al. (2006). Macroeconomic stability, fiscal policy and taxation 38 Figure 57. Production Possibility Frontier Source: World Bank Overall, results of the DEA analysis support the findings based on PSE scores (Table 8). Cro- atia has one of the lowest efficiency scores for both input and output oriented analysis. In input oriented analysis, Croatia scores 0.58 which is well below the EU10 average. This implies that Croatia could significantly reduce its expenditures and at the same time keep the same level of output. Output oriented analysis indicates that Croatia is one of the least efficient countries (rank 27 out of 28). Its score of 0.53 implies that Croatia could significantly increase the level of outputs without the need for increasing expenditures. These findings are also corroborated by other indicators: the Doing Business ranking, where it lags most of its regional peers in various indicators, and Worldwide Governance Indicators also suggest that Croatia’s public sector governance is poor and provides inadequate incentives for firms and indi- viduals. These shortcomings weigh on firms’ innovation capabilities and productivity, inhibit private sector investment in physical capital and R&D, and distort overall resource allocation. Macroeconomic stability, fiscal policy and taxation 39 Table 8. Results of Data Envelopment Analysis (variable returns to scale approach) Input oriented Output oriented Rank CRS TE Rank VRS TE Rank VRS TE Poland 2 1.00 2 1.00 2 0.73 Czech Republic 10 0.67 15 0.77 7 0.58 Slovakia 12 0.65 16 0.73 13 0.54 Slovenia 16 0.61 17 0.70 17 0.49 Estonia 11 0.65 18 0.69 16 0.51 Romania 5 0.78 19 0.69 10 0.57 Lithuania 4 0.78 20 0.66 11 0.56 Bulgaria 6 0.74 21 0.66 14 0.54 Hungary 20 0.56 23 0.64 25 0.41 Latvia 9 0.69 26 0.55 22 0.42 Croatia 18 0.58 27 0.53 26 0.35 EU 10 average 0.70 0.69 0.52 Overall average 0.65 0.75 0.53 Note: CRS TE – constant return to scale technical efficiency; VRS TE – variable returns to scale technical efficiency Source: World Bank Public sector performance in Croatia is relatively weak compared to other European coun- tries. Raising the efficiency of public sector service delivery could generate significant savings and/or boost the quality of outcomes. Strengthening outcomes is particularly important in education, and gen- eral administration (including business and product market regulation). There is also scope for improve- ments in healthcare, as well as in providing social assistance. Macroeconomic stability, fiscal policy and taxation 40 7 Conclusion While Croatia has achieved significant progress since independence, the 2008 global crisis exposed its economic weaknesses and lack of resilience, and its current growth trajectory implies a very slow pace of convergence with richer countries in the European Union. Real growth of the Croatian economy has been restored since 2015 on a more solid basis but remains one of the lowest in CEE and is expected to decelerate further in the medium run. Moreover, the current composition of the economy with a small export sector does not support strong increases in overall productivity which is critical to raise potential growth, especially in the face of an ageing population and negative migration flows. Without a step-up in productivity, Croatia could experience a deterioration of social conditions and prolonged economic stagnation or deterioration. A comprehensive reform program is therefore essential to raise productivity, create a sup- portive business environment and fostering inclusive growth. Solid macroeconomic fundamentals, decreasing unemployment and a rising demand for labor, as well as reduced fiscal sustainability con- cerns create a supportive environment for a bold reform agenda. Reforming the state both as a service provider and a regulator is of utmost importance, as public sector performance and efficiency are among the weakest in the EU. Fiscal sustainability concerns have diminished but vulnerabilities remain, and the quality of budget composition should be improved. Over the course of three years the government has achieved a strong fiscal adjustment. As fiscal imbalances narrow, and growth accelerated, public debt seems firmly on a downward path. However, at around 75 percent of GDP, a large share of which is denomi- nated in foreign currency, general government debt remains high and the economy continues to be vul- nerable to potential worsening of external borrowing conditions, interest rate shocks, and exchange rate shocks. This requires continued tight fiscal policies to accelerate debt reduction. At the same time, the composition of the budget is also a concern, given that the share of capital spending has decreased by more than half compared to the pre-crisis level, which adversely affects medium and long term growth prospects. Lastly, consideration should be given to strengthen incentives to work in order to raise labor force participation and thus mitigate the effects of ageing and emigration. In this context, the preparation of the National Development Strategy (NDS) and establish- ment of strategic planning framework creates an opportunity to streamline the reform agenda, prioritize projects with highest value added for society and make the best of the EU structural funds post-2020. Only a coordinated and guided effort, a clear prioritization of projects and regular monitoring of implementation will allow Croatia to avoid the middle-income trap and achieve sustained and inclusive growth. Macroeconomic stability, fiscal policy and taxation 41 List of references 1. Bicanic I., Generiranje postojanog rasta u Hrvatskoj u silaznoj fazi gospodarskog ciklusa, Za- greb's economic forum, 2012 2. De Kam F., Kuhry B., Pommer E., Public Sector Performance, An international comparison of education, health care, law and order and public administration, Social and Cultural Planning Of- fice, 2004 3. Drazenovic I, Kunovac M. and Pripuzic D., Dynamics and Determinants of Migration – The Case of Croatia and Experience of New EU Member States, , 2018 4. Drucker L., Geva A., What drives public sector performance?, Israel Economic Review Vol.10, No.2, 43-69, 2013 5. European Commission, Country Report Croatia 2018 Including an In-Depth Review on the pre- vention and correction of macroeconomic imbalances, Commission staff working document, 2018 6. European Commission, Taxation Trends in the European Union (2018), Annex B 7. European Commission, von Below S., PISA 2015: EU performance and initial conclusions re- garding education policies in Europe, 2016 8. European Commission, Balta N., Catching-up processes in the euro area, Quarterly report on the euro area, 2013 9. European Commission, Balta N., Mohl P., The drivers of total factor productivity in catching-up economies, Quarterly report on the euro area, 2014 10. Gabriel F., Rosenblatt D., Middle-Income Traps: A Conceptual and Empirical Survey, Policy Re- search Working Paper 6594, World Bank, 2013 11. Grela M., Majchrowska A., Michałek T, Mućk J., Stążka-Gawrysiak A., Tchorek G., Wagner M., Is Central and Eastern Europe converging towards the EU-15?, NBP Working Paper No. 264, Narodowy Bank Polski, 2017 12. Inchauste G., Rubil I., The Distributional Impact of Taxes and Social Spending in Croatia 13. Levenko N., Oja K., Staehr K., Total factor productivity growth in Central and Eastern Europe before, during and after the global financial crisis, Eesti Pank. Working Paper Series, 2017 14. Melitz J. M., Polanec S., Dynamic Olley-Pakes productivity decomposition with entry and exit, Working Paper 18182, NBER, 2012 15. Orsini K, Ostojic V., Croatia's Tourism Industry: Beyond the Sun and Sea, EC, 2018 16. Orsini K, What drives Croatia's high import dependence?, European Commission, 2017 17. Perusko V.I., Kovac K., M. Josic, Croatia in Global Value Chains, Surveys S-32, CNB, 2018 18. Rapacki R., Prochniak M., The EU enlargment and Economic Growth in the CEE New Member Countries, Economic Papers 367, European Commission, 2009 19. Ranilovic, N., The Effects of Economic Integration on Croatian Merchandise Trade: A Gravity Model Study, Working Paper W-50, Croatian National Bank, 2017 20. Staehr K., Economic Growth and Convergence in the Baltic States: Caught in a Middle Income Trap?, DG ECFIN seminar, Tallinn University of Technology, 2015 21. Valdec M, Zrnc J., The direction of causality between exports and firm performance: microeco- nomic evidence from Croatia using the matching approach, 2014 22. van Ark, B., Total factor productivity: Lessons from the past and directions for the future, The Conference Board presentation, 2014 Macroeconomic stability, fiscal policy and taxation 42 Annex I. Public sector performance and efficiency: varia- bles and methodology 1 Variables and series Table 1. Variables and series Indicator: Variable: Series: Source: The Global Competitiveness Index Historical Dataset, 2007-2017 World Corruption index Ethics and corruption, 1-7 (best) Economic Forum The Global Competitiveness Index Historical Dataset, 2007-2017 World Administrative Red tape Burden of government regulation, 1-7 (best) Economic Forum Efficiency of legal framework in settling The Global Competitiveness Index Historical Dataset, 2009-2017 World Efficient judiciary disputes, 1-7 (best) Economic Forum Schneider, F. (2015), "Size and Development of the Shadow Economy of 31 Size shadow economy Shadow Economy in % of GDP European and 5 other OECD Countries from 2003 to 2015" Secondary school enrollment School enrollment, secondary (% gross) World development indicators, World Bank (up to 2016) Education PISA report -Simple average of reading, Education achievement mathematics and science scores PISA 2015 results, OECD Health Infant mortality rate Mortality rate, infant (per 1,000 live births) World development indicators, World Bank (up to 2016) Life expectancy Life expectancy at birth, years World development indicators, World Bank (up to 2016) Quality of The Global Competitiveness Index Historical Dataset, 2007-2017 World infrastructure Quality of infrastructure Quality of overall infrastructure, 1-7 (best) Economic Forum Distribution Income distribution GINI Eurostat Gross domestic product, chain linked Coefficient of variation of growth volumes Eurostat Stability Standard deviation of inflation/inflation average HICP Eurostat GDP per capita in 2017 US$ (converted to The Conference Board Total Economy Database™ (Adjusted version), March Per capita income 2017 price level with updated 2011 PPPs) 2018 Economic Gross domestic product, chain linked performance Average economic growth volumes Eurostat Unemployment rate Unemployment rate/ 10 year average Eurostat General government total expenditure in % Total public expenditure of GDP Eurostat Intermediate consumption / General government intermediate government consumption consumption in % of GDP Eurostat General government education expenditure Expenditure Public education in % of GDP - COFOG Eurostat categories General government health expenditure in % Public health of GDP - COFOG Eurostat General government social transfers other Transfers than social transfers in kind in % of GDP Eurostat General government gross fixed capital Public investment formation in % of GDP Eurostat 2 Methodology and data We use three approaches in this study. The first section explains how public sector performance indica- tor is constructed while second and third section provide explanation of approaches used to assess public sector efficiency. 2.1 Public sector performance Following Afonso et al. (2005) public sector performance is assessed by constructing composite indica- tors which are based on socio-economic indicators that are assumed to be the output of public policies. We can write this as: Macroeconomic stability, fiscal policy and taxation 43 = ∑ , i = 1, … , n: with = ( ), = 1, … , (1) =1 where ( ) is a function of k observable selected variables. Indicators are divided into two groups as shown by Figure 1. The first group includes outcomes of public policies aimed at providing various and accessible public services for individuals. Afonso et al. describe these as opportunity indicators, referring to the role of the government as a promoter of equal opportunities in the market place. The second group include indicators linked to the standard Musgra- vian functions of the government - allocation, distribution and stabilization. Opportunity indicators are divided into four sub-groups which reflect government performance in administration, education, health and infrastructure. The administration sub-indicator is comprised of indicators measuring corruption, burden of regulation, judiciary quality and shadow economy. For the education sub-indicator we included data on secondary school enrolment rate and PISA scores in order to measure both education quantity and quality. The health performance indicator contains data on infant mortality and life expectancy while the infrastructure sub-indicator is measured by the quality of overall infrastructure. When it comes to standard Musgravian indicators, income distribution is measured by the Gini index. For the stability sub-indicator we used coefficient of variation of GDP growth and standard de- viation of inflation (10 year average, i.e. 2008 – 2017 period). Data used in this analysis is compiled from various sources that are listed in Table 1 in the Appen- dix. In the case of the economic variables we use 10-years averages given that they vary strongly from one year to the next. The rest of the variables correspond to 2017 or closest available year as they change very slowly over time. Macroeconomic stability, fiscal policy and taxation 44 Figure 1. Public Sector Performance Indicator PSP Corruption Distribution Gini Index Red Tape Administrative Quality of judiciary Shadow economy Stability of GDP growth (coefficient of Secondary variation) school Stability enrollment Standard Education deviation of Education inflation achievement GDP per capita Infant mortality Economic GDP growth Health Performance Life expectancy Unempoloyment Quality of Public infrastructure Infrastructure Sources: Afonso et al. (2005, 2006) The PSP indicator for each country is obtained by aggregation of sub-indicators after assigning equal weights to them and each sub-indicator is computed as the simple average of the corresponding variables. In order to facilitate the comparison, all the variables are first normalized by dividing the value of specific country with the average of that measure for all the countries in the sample. 2.2 Public sector efficiency Public sector efficiency indicator is constructed by relating public sector performance indicators with relevant general government expenditures. For each country i with j areas of public sector performance Public sector efficiency indicator can be written as: = ∑ =1 , = 1, … , (2) Government performance in the area of administration is weighted by government collective con- sumption. Education and health performance are weighted by general government spending in education Macroeconomic stability, fiscal policy and taxation 45 and health, respectively, while quality of infrastructure is weighted by general government investment expenditures. As for the Musgravian functions, distribution performance is weighted by spending on social transfers other than social transfers in kind, while stability and economic performance were re- lated to total general government spending. General government spending was normalized across countries with the average taking the value of one for each specified sub-indicator. Different categories of general government expenditures are computed as 10-year averages for the period 2008-2017 given that we assume a lagged effect of spend- ing to government performance. 2.3 Data envelopment analysis As noted by Afonso et. al (2006) the DEA methodology assumes the existence of a convex production frontier where production frontier is constructed by using linear programming methods. The term “en- velopment” stems from the fact that the production frontier envelops the set of observations. DEA can be conducted for the input and output-oriented analysis. Input oriented specification shows how much input quantities can be proportionally reduced without changing the output quantities. Conversely, output oriented specification tries to assess how much output can be proportionally in- creased without changing the input quantities used. The two measures provide the same results under constant returns to scale but give different values under variable returns to scale. For detailed analytical description see Afonso et al. (2006). An example of DEA analysis is illustrated on Figure 2. The y axis represents the output while x axis the input used by the four countries. Under variable returns to scale assumption, frontier is deter- mined by country A and C and countries B and D can therefore be considered as inefficient. They have worse status although their expenses are higher. Under constant return to scale assumption we can see that only country A can be considered as efficient. Figure 2. Example of DEA frontiers Source: Afonso et al. (2006) Macroeconomic stability, fiscal policy and taxation 46 Annex II. Debt Sustainability Analysis Assumptions Individual shocks are permanent and are defined in proportion to standard deviation calculated for 2008- 2018 period. First year of forecast is 2019. Key macroeconomic and fiscal assumptions 2018 2019 2020 2021 2022 2023 Baseline scenario Real GDP growth, % 2.6 2.5 2.5 2.4 2.4 2.4 Average nominal interest rate on public debt, % 3.1 3.0 3.0 3.0 3.0 2.9 General government balance, % of GDP 0.2 0.1 0.2 0.4 0.4 0.4 General government primary balance, % of GDP 1.7 2.5 2.4 2.5 2.4 2.3 Primary (noninterest) expenditure, % of GDP 44.1 43.0 42.6 42.2 41.8 41.4 Exchange rate (HRK/EUR) 7.4 7.4 7.4 7.4 7.4 7.4 Interest rate shock scenario Previous year value + 2 standard deviation, 10-year value Average nominal interest rate on public debt, % 3.1 5.6 5.6 5.7 5.6 5.5 GDP growth shock scenario Prevous year value + 1 standard deviation, 10-year value Real GDP growth, % 2.6 -0.7 -0.7 -0.8 -0.8 -0.8 Primary balance shock scenario Prevous year value + 1 standard deviation, 10-year value General government balance, % of GDP 0.2 -2.6 -2.5 -2.2 -2.3 -2.3 General government primary balance, % of GDP 1.7 -0.2 -0.3 -0.2 -0.3 -0.4 Primary (noninterest) expenditure, % of GDP 44.1 45.7 45.3 44.9 44.5 44.1 Contingent liabilities shock scenario Materialization of contingent liabilities in amount of 50% State guarantees in the first year of projection Recognition of implicit or contingent liabilities, % of GDP 0.0 1.5 0.0 0.0 0.0 0.0 Depreciation shock scenario 20% depreciation of HRK towards EUR in the first year of projection Exchange rate (HRK/EUR) 7.4 8.9 8.9 8.9 8.9 8.9 Combined shock scenario GDP growth shock + Depreciation shock + Interest rate shock (with added 0.5 standard deviation and 15% depreciation of HRK towards EUR in the first year of projection) Real GDP growth, % 2.6 0.9 0.9 0.8 0.8 0.8 Average nominal interest rate on public debt, % 3.1 3.7 3.7 3.7 3.6 3.6 Exchange rate (HRK/EUR) 7.4 8.5 8.5 8.5 8.5 8.5 Source: World Bank Staff calculation. Content of this report is the sole responsibility of the World Bank. Macroeconomic stability, fiscal policy and taxation 47